BGPT: Paper Review: Thermodynamics of amyloid fibril formation from chemical depolymerization

Explore by Goal

Quick Explanation Copied

Paper verdict (skeptical + evidence-weighted)

The study shows that fitting amyloid depolymerization with a single-constant isodesmic linear polymer model can yield concentration-dependent artifacts, while a cooperative (nucleation–elongation) model better captures the thermodynamic data for glucagon and PI3K-SH3 and enables an ionic-strength stability analysis consistent with a product-like transition state for elongation.

Key support: the isodesmic model yields different extrapolated ΔG° values when peptide concentration changes, whereas the cooperative model fits more consistently and yields specific parameters (small effective nucleus size n≈2–5) and ionic-strength trends.

Main paper:

Long Explanation

Thermodynamics of amyloid fibril formation from chemical depolymerization — rigorous review

Context: Amyloid fibril lifetime/clearance is influenced by thermodynamic stability, but standard depolymerization analyses often use linear polymerization assumptions (commonly isodesmic, single-constant models) that may be thermodynamically inadequate for nucleated polymerization.

1) Visual data extraction (from the provided paper text)

The review below focuses on (i) the concentration dependence problem that exposes model misspecification, (ii) the cooperative model’s recovery of consistency, and (iii) ionic-strength trends used to infer the nature of the elongation transition state.

Figure A — Isodesmic model yields concentration-dependent ΔG° (artifact)

Interpretation (skeptical): The provided extracted values in the paper text indicate that fitting each glucagon peptide concentration separately with the isodesmic model yields different inferred ΔG° values (−36.7 vs −38.8 vs −42.2 kJ/mol). For PI3K-SH3, a similar concentration-dependent spread appears (−64.7 vs −71.1 kJ/mol).

Figure B — Cooperative model collapses the concentration dependence into consistent ΔG°

Interpretation: The paper text explains that global isodesmic fits can fail (especially for glucagon) while cooperative model fits (with parameters for nucleation vs elongation) better account for the concentration dependence; the cooperative model also performs better when peptide concentration is varied (a crucial experimental discriminator).

2) Core methodological logic (what they actually measured)

Systems: two amyloid-forming polypeptides under acidic conditions: human glucagon and bovine PI3K-SH3 domain.
Readout: intrinsic Trp fluorescence ratios are used as a proxy for the fraction soluble vs fibrillar after chemical depolymerization, using specific excitation/emission settings and wavelength ratios chosen by differences between fibrillar and monomeric spectra.
Equilibration: samples are equilibrated for long periods (reported as one week for glucagon depolymerization series, two weeks for PI3K-SH3) to support equilibrium assumptions.

3) Model comparison: isodesmic vs cooperative nucleation–elongation

Isodesmic linear polymerization: assumes a single equilibrium constant for monomer association across aggregate sizes, which the paper argues is inconsistent with nucleated polymerization features of amyloid formation.

Cooperative model: adapted from supramolecular polymerization; introduces two equilibrium constants: one for monomer association up to a nucleus size n and one for elongation beyond it, with an effective ratio s = k_n/k_e.

Why varying peptide concentration matters: The paper emphasizes that standard experiments vary denaturant concentration at fixed peptide concentration, while varying peptide concentration provides a stronger discrimination between isodesmic and cooperative model adequacy.

Result they claim: best cooperative fits are achieved for small to intermediate nucleus sizes (n≈2–5) and relatively small s (i.e., nucleation less favorable than elongation).

4) Ionic strength: stability shifts and an inference about the elongation transition state

Core observation: increasing NaCl stabilizes both amyloid fibrils under the study’s acidic conditions, shifting depolymerization midpoints to higher denaturant concentrations.

Quantification: they fit the depolymerization curves using fixed cooperative parameters (from best fits) and then plot log equilibrium constants vs square root of ionic strength; they report slopes/intercepts for PI3K-SH3 and glucagon.

Figure C — Ionic-strength sensitivity slopes (reported)

Skeptical reading: The paper interprets higher ionic-strength slope for PI3K-SH3 (vs glucagon) as greater sensitivity consistent with higher net positive charge at pH 2. This is plausible given charge screening arguments, but the paper’s inference depends on the specific theoretical mapping used for log(K) vs √I and on whether denaturant/urea conditions preserve the relevant electrostatic regime.

5) Product-like transition state: thermodynamics vs kinetics comparison

The paper compares salt-dependence slopes for thermodynamic stability (this study) with previous kinetic elongation data (ref. 27 in paper text) and argues that similar slopes imply that the elongation transition state ensemble is highly product-like (i.e., the ion-screening signature seen in kinetics resembles that of the final state incorporation geometry).

Figure D — Kinetics vs thermodynamics slope comparison (reported)

Where this could be misleading: “Slope similarity” is informative but not a complete mechanistic proof: it depends on shared theoretical assumptions linking ionic strength to electrostatic free-energy contributions in both kinetic and equilibrium analyses. The paper explicitly notes limited overlap in ionic strength ranges due to experimental buffering constraints, and compares slopes semi-quantitatively for selected points.

6) Evidence strengthening: equilibrium vs kinetics; alternative thermodynamic probes

Because depolymerization equilibrium inference is nontrivial, it’s useful to triangulate with independent thermodynamic measurement approaches. For example, ITC-based thermodynamic parameterization of amyloid extension (β2-microglobulin) supports the idea that amyloid formation can show distinct enthalpy/heat-capacity behavior (including entropy-driven contributions under certain conditions), reinforcing that thermodynamics is experimentally measurable rather than purely model-based.

7) Critical appraisal (skeptical blind spots + what would disprove the key claims)

Claim 1: Isodesmic model is inadequate across concentration variations.

Strength: The paper uses a concentration-dimension discriminator and reports explicit concentration-dependent discrepancies in fitted ΔG° for glucagon (and similarly for PI3K-SH3) under the isodesmic model.
Blind spots: Because ΔG° and the denaturant slope m are both free parameters, misspecification can manifest as parameter drift; this is partly handled by global fitting, but the success of cooperative fitting could still reflect the flexibility introduced by additional parameters (overfitting risk). The paper partially mitigates this by fitting via a physically motivated cooperative nucleation model and testing multiple nucleus sizes n.

Claim 2: Cooperative polymerization with small effective nucleus size (n≈2–5) describes thermodynamics.

Strength: The paper reports systematic model extension and comparison, including alternative assumptions about the fluorescence signature of oligomers (osaa vs osam) and reports that osaa yields more consistent ΔG° and s values.
Blind spots / known-unknowns: The “effective nucleus size” is a model parameter, not a directly observed physical nucleus. The paper itself notes amyloid formation involves multiple steps (monomer addition and structural rearrangement), so the cooperative model is likely an oversimplification—yet it may capture the essence of thermodynamic behavior.

Claim 3: Ionic-strength dependence supports a product-like transition state ensemble for elongation.

Strength: The paper connects the thermodynamic ionic-strength slope to previously published kinetics and argues similar slopes imply proximity of the transition state to the product geometry (late/product-like).
Disproof path: If later work shows that ionic-strength signatures in equilibrium and kinetics correspond to different physical distances/charge configurations, slope similarity could be coincidental. The strongest disproof would be a systematic mismatch when changing salt type/ionic strength mapping or when directly measuring transition-state electrostatics rather than inferring from slope relations.

8) Author-funding / conflicts (sanity checks)

Funding is listed as DFG and Novo Nordisk entities, and the paper states no conflicts of interest.

Critical note: Funding disclosure and COI statements do not guarantee absence of bias, but they reduce the chance of undisclosed conflicts.

9) Buttons for deeper BGPT exploration

10) Run an iterative Science AI agent (optional)

Author reviews (tap-through)

Feedback:

Updated: April 13, 2026